From: guido AT sirio DOT stat DOT unipd DOT it (Guido Masarotto)
Subject: Mingw32, g77 and B20
9 Nov 1998 01:50:00 -0800
Message-ID: <19981108193344.B2044.cygnus.gnu-win32@sirio.stat.unipd.it>
References: <Pine DOT SUN DOT 3 DOT 93 DOT 981106163910 DOT 10108B-100000 AT modi DOT xraylith DOT wisc DOT edu>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
To: gnu-win32 AT cygnus DOT com

I have recently ported a statistical program (R = "Gnu S") using
egcs-mingw32 (http://www.ci.tuwien.ac.at/R/contents.html).

R require both C and Fortran. Today, I tried to recompile it
using B20 + the g77 add-on made available by Mumit Khan.
I used the mingw32 flag (-mno-cygwin) after adjusting `specs'
(as discussed by many messages in these days).
Compilation was ok but linking gave a lot of unresolved symbols.
The problem is that the Fortran library in the Mumit's add-on
(libg2c) is for Cygwin not for Mingw32.
Next, I linked not against the B20 libg2c but against the
egcs-1.1-mingw32 libg2c (theory: it is the same version of egcs,
the libraries must be the same, this last one is built for 
Mingw32). Linking was okay and R seems to work ok
(also some dll's (mixing Fortran and C) dyn.loaded by R at runtime).

Question is: is this reliable? Or, what must be done to use
Fortran with Mingw32 and B20?

Thanks in advance,
guido 




-
For help on using this list (especially unsubscribing), send a message to
"gnu-win32-request AT cygnus DOT com" with one line of text: "help".